Machine learning feature extraction in naturalistic stimuli for human brain mapping using high-density diffuse optical tomography

Fogarty, Morgan; Tripathy, Kalyan; Svoboda, Alexandra M.; Schroeder, Mariel L.; Rafferty, Sean M.; Mansfield, Patricia; Ulbrich, Rachel L.; Booth, Madison; Richter, Edward J.; Smyser, Christopher D.; Eggebrecht, Adam T.; Culver, Joseph P.

doi:10.1117/12.2608946

Cited by 1 publication

(1 citation statement)

References 0 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Recent research has attempted to address some of the challenges posed due to the diffuse nature of photons and ill-posed inverse problems by employing deep learning (DL) [ 15 , 17 , 20 , 25 , 26 ] and shallow machine learning (ML) techniques [ 8 , 27 ]. The usage of ML algorithms regularly beats the analytical solution technique in all of the articles previously mentioned.…”

Section: Introductionmentioning

confidence: 99%

Machine Learning Diffuse Optical Tomography Using Extreme Gradient Boosting and Genetic Programming

2023

View full text Add to dashboard Cite

Diffuse optical tomography (DOT) is a non-invasive method for detecting breast cancer; however, it struggles to produce high-quality images due to the complexity of scattered light and the limitations of traditional image reconstruction algorithms. These algorithms can be affected by boundary conditions and have a low imaging accuracy, a shallow imaging depth, a long computation time, and a high signal-to-noise ratio. However, machine learning can potentially improve the performance of DOT by being better equipped to solve inverse problems, perform regression, classify medical images, and reconstruct biomedical images. In this study, we utilized a machine learning model called “XGBoost” to detect tumors in inhomogeneous breasts and applied a post-processing technique based on genetic programming to improve accuracy. The proposed algorithm was tested using simulated DOT measurements from complex inhomogeneous breasts and evaluated using the cosine similarity metrics and root mean square error loss. The results showed that the use of XGBoost and genetic programming in DOT could lead to more accurate and non-invasive detection of tumors in inhomogeneous breasts compared to traditional methods, with the reconstructed breasts having an average cosine similarity of more than 0.97 ± 0.07 and average root mean square error of around 0.1270 ± 0.0031 compared to the ground truth.

show abstract